[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6ff6fdea-c955-f9dd-289e-b0d613a28280@I-love.SAKURA.ne.jp>
Date: Sat, 22 Apr 2023 00:12:02 +0900
From: Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>
To: Dmitry Vyukov <dvyukov@...gle.com>,
syzbot <syzbot+b7c3ba8cdc2f6cf83c21@...kaller.appspotmail.com>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Jiri Slaby <jirislaby@...nel.org>
Cc: linux-kernel@...r.kernel.org, syzkaller-bugs@...glegroups.com
Subject: Re: [syzbot] [kernel?] KCSAN: data-race in __fput / __tty_hangup (4)
On 2023/04/21 17:21, Dmitry Vyukov wrote:
> If I am reading this correctly, this race can lead to NULL derefs
> among other things.
> hung_up_tty_fops does not have splice_read, while other fops have.
>
> So the following code in splice can execute NULL callback:
>
> if (unlikely(!in->f_op->splice_read))
> return warn_unsupported(in, "read");
> return in->f_op->splice_read(in, ppos, pipe, len, flags);
>
__fput(file) is called when the last reference to file is released.
Since __tty_hangup() traverses tty->tty_files under tty->files_lock,
tty_add_file() needs to hold a ref before adding to tty->tty_files
in order to defer concurrent __fput() by other threads?
diff --git a/drivers/tty/tty_io.c b/drivers/tty/tty_io.c
index 36fb945fdad4..2838703d48cf 100644
--- a/drivers/tty/tty_io.c
+++ b/drivers/tty/tty_io.c
@@ -197,7 +197,7 @@ void tty_add_file(struct tty_struct *tty, struct file *file)
struct tty_file_private *priv = file->private_data;
priv->tty = tty;
- priv->file = file;
+ priv->file = get_file(file);
spin_lock(&tty->files_lock);
list_add(&priv->list, &tty->tty_files);
@@ -228,6 +228,7 @@ static void tty_del_file(struct file *file)
spin_lock(&tty->files_lock);
list_del(&priv->list);
spin_unlock(&tty->files_lock);
+ fput(file);
tty_free_file(file);
}
Powered by blists - more mailing lists